14  Diverging bar graphs

This graph is largely complete and just needs final proof reading.

This graph requires:

✅ a categorical variable

✅ a numeric (continuous) variable

14.1 Description

If you have two proportions that contain positive and negative values, consider using diverging bars with geom_bar().

Unlike a standard or stacked bar graphs, diverging bar graphs display positive and negative quantities on both sides of a reference or baseline value (zero in this example). Color, length and position are used to compare the quantities across categorical levels (and within variable values).

14.1.1 Horizontal bars

For example, we can use the length of the bar from the reference line to compare disapproval estimates across all months (i.e., comparing red bars to each other).

14.1.2 Vertical bars

We can also compare approval vs. disapproval for each month (i.e., compare the blue vs. red bars to each other within each month).

14.2 Set up

PACKAGES:

Install packages.

install.packages("fivethirtyeight")
library(fivethirtyeight) 
library(ggplot2)

DATA:

Create trump_approval_diverg from the trump_approval_trend dataset in the fivethirtyeight package.

fivethirtyeight::trump_approval_trend |> 
  dplyr::filter(subgroup == "All polls") |> 
  dplyr::mutate(
    month = lubridate::month(modeldate,
                label = TRUE, abbr = TRUE),
    approve = approve_estimate*0.01,
    disapprove = disapprove_estimate*0.01,
    disapprove = disapprove * -1) |> 
  tidyr::pivot_longer(cols = c(approve, disapprove), 
    names_to = "poll", values_to = "values") |> 
  dplyr::group_by(month, poll) |> 
    dplyr::summarise(
      month_avg = mean(values, na.rm = TRUE)
    ) |> 
  dplyr::ungroup() -> trump_approval_diverg
#> `summarise()` has grouped output by 'month'. You
#> can override using the `.groups` argument.
glimpse(trump_approval_diverg)
#> Rows: 24
#> Columns: 3
#> $ month     <ord> Jan, Jan, Feb, Feb, Mar, Mar, …
#> $ poll      <chr> "approve", "disapprove", "appr…
#> $ month_avg <dbl> 0.4029758, -0.5234634, 0.42260…

14.3 Grammar

CODE:

• Create labels with labs()

• Initialize the graph with ggplot() and provide data

• Map the month to the x and month_avg to the y

• Insidegeom_bar()

  • map poll to fill

  • use stat = "identity" and width = .5

• Add scale_y_continuous() to manually set the limits and format the axis with scales::percent

labs_geom_bar_diverg <- labs(
  title = "Trump Approval Ratings",
  subtitle = "From 'How Popular is Donald Trump'",
  x = "Month",
  y = "Monthly average percent",
  fill = "Estimate")
ggp2_bars_diverg <- ggplot(
  data = trump_approval_diverg, 
    aes(x = month, y = month_avg)) +
  geom_bar(aes(fill = poll),
    stat = "identity", width = .5) + 
  scale_y_continuous(limits = c(-1, 1), 
    labels = scales::percent) 

ggp2_bars_diverg + 
  labs_geom_bar_diverg

GRAPH:

14.4 More info

For vertically arranged bars, we switch the x and y axis variables.

14.4.1 Vertically arranged bars

• Create labels with labs()

• Map the month_avg to the x and month to the y

• Inside geom_bar()

  • map poll to fill

  • use stat = "identity" and width = .5

• Add scale_y_continuous() to manually set the limits and format the axis with scales::percent

labs_geom_bar_diverg_vert <- labs(
  title = "Trump Approval Ratings",
  subtitle = "From 'How Popular is Donald Trump'",
  x = "Monthly average percent",
  y = "Month",
  fill = "Estimate")
ggp2_bar_diverg_vert <- ggplot(
  data = trump_approval_diverg,
      aes(x = month_avg, y =  month)) +
  geom_bar(
      aes(fill = poll), 
        stat = "identity", width = .5) + 
  scale_x_continuous(limits = c(-1, 1), 
        labels = scales::percent)
ggp2_bar_diverg_vert + 
  labs_geom_bar_diverg_vert